AIQF Daily digest — 11 new papers

Quant Paper Radar · 6 of 11

1. 1

   Portfolio Optimization Under Heavy Tails and Asymmetric Volatility: Evidence from Taiwan-Exposed ETFs

   Ting-Jung Lee, Abootaleb Shirvani, Farzana Afroz et al. · OPENALEX · Tail-risk portfolio optimization

   we backtested it — Sharpe 0.37 · quant relevance 0.96

   The study compares mean–variance and CVaR allocations for U.S.-listed ETFs exposed to Taiwan and semiconductors, using daily returns to assess tail risk and asymmetric volatility. It ranks the portfolios by historical performance, though the concentrated CVaR allocations may depend on the sample.

   original paper

2. 2

   AlphaPADI: Formulaic Alpha Discovery via Pool-Aware Hierarchical Discrete Diffusion

   Yanzheng Jin, Pengyang Shao, Yunshan Ma et al. · ARXIV · Machine-Learned Alpha Discovery

   quant relevance 1.00

   AlphaPADI builds pools of complementary symbolic signals to predict cross-sectional stock returns. The authors report improved predictive and portfolio performance in Chinese and U.S. equities, and the U.S. component can be tested with the available data.

   original paper

3. 3

   A Price-Based Framework for Stochastic Portfolio Theory

   Jongbong An, Donghan Kim · ARXIV · Functional portfolio generation

   we backtested it — Sharpe 0.38 · quant relevance 0.94

   The paper develops price-weight-based equity portfolios that explicitly account for stock splits and compares their measured performance with capitalization-weight-based portfolios on NYSE data. Its rules have a direct investing application, but the full 2002–2021 sample cannot be reproduced.

   original paper

The rest of the batch

• FactorBench: A Portfolio-Aware Benchmark for Automated Factor Mining

  Zhuohan Wang, Carmine Ventre · ARXIV · relevance 0.98

  FactorBench tests whether automatically mined equity signals generalize and improve long-only and long–short portfolios after costs.

• Deep Learning-Based Gold Price Prediction and Cross-Market Arbitrage Strategy with Currency Hedging

  Zijian Zeng, Fei Ding, Yi Wensheng · OPENALEX · relevance 0.98

  The paper tests a daily SHFE–COMEX gold futures spread trade filtered by a model and hedged with SGX USD/CNH futures.

• Adversarial Training for Deep Hedging in Nonstationary Markets

  Philipp J. Schneider, Lukas Looser, Antoine Garin et al. · ARXIV · relevance 0.96

  The paper proposes adversarial training for dynamic derivative hedges under changing market conditions.

and 5 more · see the full radar